U.S. patent application number 11/854217 was filed with the patent office on 2009-03-12 for method and apparatus for detecting anomalies in landing systems utilizing a global navigation satellite system.
Invention is credited to Tiffany R. Lapp, Timothy Allen Murphy.
Application Number | 20090069960 11/854217 |
Document ID | / |
Family ID | 40172017 |
Filed Date | 2009-03-12 |
United States Patent
Application |
20090069960 |
Kind Code |
A1 |
Lapp; Tiffany R. ; et
al. |
March 12, 2009 |
METHOD AND APPARATUS FOR DETECTING ANOMALIES IN LANDING SYSTEMS
UTILIZING A GLOBAL NAVIGATION SATELLITE SYSTEM
Abstract
A method, apparatus, and computer program product for detecting
anomalies in a landing system. In one embodiment, a magnitude
difference between a blended inertial deviation magnitude and a raw
deviation magnitude is identified to form a magnitude difference.
The magnitude difference is compared to a magnitude threshold. If
the magnitude difference exceeds the magnitude threshold, an
anomaly in the data is detected.
Inventors: |
Lapp; Tiffany R.;
(Snohomish, WA) ; Murphy; Timothy Allen; (Everett,
WA) |
Correspondence
Address: |
DUKE W. YEE
YEE & ASSOCIATES, P.C., P.O. BOX 802333
DALLAS
TX
75380
US
|
Family ID: |
40172017 |
Appl. No.: |
11/854217 |
Filed: |
September 12, 2007 |
Current U.S.
Class: |
701/16 ;
340/947 |
Current CPC
Class: |
G05D 1/0077
20130101 |
Class at
Publication: |
701/16 ;
340/947 |
International
Class: |
G06F 19/00 20060101
G06F019/00 |
Claims
1. A computer implemented method for detecting data anomalies in a
landing system, the computer implemented method comprising:
identifying a magnitude difference between a blended inertial
deviation magnitude and a raw deviation magnitude in a landing
system associated with a global navigation satellite system to form
a magnitude difference; comparing the magnitude difference to a
magnitude threshold; and responsive to the magnitude difference
exceeding the magnitude threshold, detecting an anomaly in the
data.
2. The computer implemented method of claim 1 further comprising:
measuring a difference between a blended inertial deviation rate
and a raw deviation rate in the landing system associated with the
global navigation satellite system to form a rate difference;
responsive to the rate difference exceeding the magnitude
threshold, determining whether the rate difference indicates a ramp
or a step; and responsive to the rate difference indicating a ramp
error, detecting an anomaly in the data.
3. The computer implemented method of claim 1 further comprising:
generating a dynamic inertial error estimate threshold; and
comparing an un-lagged bias estimate to the dynamic inertial error
estimate threshold to determine whether a ramp error is
detected.
4. The computer implemented method of claim 3 wherein the dynamic
inertial error estimate threshold comprises an upper threshold and
a lower threshold, and further comprising: responsive to the
un-lagged bias estimate exceeding the upper dynamic threshold,
detecting the ramp error; and responsive to the un-lagged bias
estimate being lower than the lower threshold, detecting the ramp
error.
5. The computer implemented method of claim 3 wherein generating
the dynamic threshold further comprises: inputting a bias estimate
into a lag filter and generating an upper bound and a lower bound
around the bias estimate; generating an upper dynamic threshold
using the upper bound; and generating a lower dynamic threshold
using the lower bound, wherein the upper dynamic threshold and the
lower dynamic threshold forms the dynamic threshold.
6. The computer implemented method of claim 2 further comprising:
responsive to detecting an anomaly, triggering inertial
coasting.
7. The computer implemented method of claim 4 further comprising:
responsive to detecting the ramp error, indicating a loss of
extended coasting capability associated with the global navigation
satellite system.
8. The computer implemented method of claim 5 wherein the bias
estimate is a velocity bias estimate.
9. The computer implemented method of claim 5 wherein the bias
estimate is an acceleration bias estimate.
10. A computer program product comprising: a computer usable medium
including computer usable program code for detecting data anomalies
in a landing system, the computer program product comprising:
identifying a magnitude difference between a blended inertial
deviation magnitude and a raw deviation magnitude in a landing
system associated with a global navigation satellite system to form
a magnitude difference; comparing the magnitude difference to a
magnitude threshold; and responsive to the magnitude difference
exceeding the magnitude threshold, triggering inertial
coasting.
11. The computer program product of claim 10 further comprising:
computer usable program code for measuring a difference between a
blended inertial deviation rate and a raw deviation rate in the
landing system associated with the global navigation satellite
system to form a rate difference; computer usable program code for
determining whether the rate difference indicates a ramp or a step
in response to the rate difference exceeding the magnitude
threshold in response to the rate difference exceeding the
magnitude threshold; and computer usable program code for detecting
an anomaly in the data in response to the rate difference
indicating a ramp.
12. The computer program product of claim 10 further comprising:
computer usable program code for generating a dynamic inertial
error estimate threshold; and computer usable program code for
comparing an un-lagged bias estimate to the dynamic inertial error
estimate threshold to determine whether a ramp error is
detected.
13. The computer program product of claim 12 wherein the dynamic
inertial error estimate threshold comprises an upper threshold and
a lower threshold, and further comprising: computer usable program
code for detecting the ramp error in response to the un-lagged bias
estimate exceeding the upper dynamic threshold; and computer usable
program code for detecting the ramp error in response to the
un-lagged bias estimate being lower than the lower threshold.
14. The computer program product of claim 12 further comprising:
computer usable program code for inputting a bias estimate into a
lag filter; computer usable program code for biasing the bias
estimate in a positive direction and in a negative direction to
generate an upper bound and a lower bound around the bias estimate;
computer usable program code for generating an upper dynamic
threshold using the upper bound; and computer usable program code
for generating a lower dynamic threshold using the lower bound,
wherein the upper dynamic threshold and the lower dynamic threshold
forms the dynamic threshold.
15. The computer program product of claim 10 further comprising:
computer usable program code for triggering inertial coasting in
response to detecting an anomaly.
16. The computer program product of claim 13 further comprising:
computer usable program code for indicating a loss of extended
coasting capability associated with the global navigation satellite
system in response to detecting the ramp error.
17. An anomaly detector in a landing system associated with a
global navigation satellite system, the anomaly detector
comprising: a coasting filter, wherein the coasting filter
generates a blended inertial deviation magnitude and a blended
inertial deviation rate; a coast skip reset trigger anomaly
detector, wherein the coast skip reset trigger anomaly detector
identifies a magnitude difference between the blended inertial
deviation magnitude and a raw deviation magnitude to form a
magnitude difference; measures a difference between a blended
inertial deviation rate and a raw deviation rate to form a rate
difference; compares the magnitude difference to a magnitude
threshold; detects an anomaly in the data in response to the
magnitude difference exceeding the magnitude threshold, detecting
an anomaly in the data; and detects the anomaly in the data if the
rate difference exceeds the magnitude threshold and the rate
difference indicates a ramp.
18. The anomaly detector of claim 17 further comprising: a slow
ramp detector, wherein the slow ramp detector generates a dynamic
inertial error estimate threshold; and compares an un-lagged bias
estimate to the dynamic inertial error estimate threshold to
determine whether a ramp error is detected.
19. The anomaly detector of claim 18 further comprising: a lag
filter, wherein the lag filter receives a bias estimate; biases the
bias estimate in a positive direction and in a negative direction
to generate an upper bound and a lower bound around the bias
estimate; and wherein the slow ramp detector generates an upper
dynamic threshold using the upper bound; and generates a lower
dynamic threshold using the lower bound, wherein the upper dynamic
threshold and the lower dynamic threshold forms the dynamic
threshold.
20. The anomaly detector of claim 17 further comprising: an
indicator, wherein the indicator indicates a loss of extended
coasting capability associated with the global navigation satellite
system in response to detecting a ramp error.
Description
BACKGROUND INFORMATION
[0001] 1. Field
[0002] The present disclosure relates generally to processing data,
in particular, to landing systems. Still more particularly, the
present disclosure relates to a method, apparatus, and computer
usable program code for detecting data anomalies in a landing
system utilizing a global navigation satellite system.
[0003] 2. Background
[0004] An autopilot is a system used to guide a vehicle, such as an
airplane, with little or no intervention by a human. Autopilots
typically rely on signals transmitted from a ground based station.
These signals are used to determine the position of the vehicle
with respect to other objects, such as the runway. The autopilot
reads its position from a guidance system, such as an instrument
landing system (ILS). The autopilot uses error reduction systems to
identify and dissipate errors in the navigation information. Errors
may occur due to problems such as loss of signal, beam bends,
noise, multi-pathing, and oscillatory behavior occurring during
over flight interference or fly-through events.
[0005] Global navigation satellite system (GNSS) based landing
systems (GLS) for aircraft are becoming more widespread as they
offer improved accuracy in navigation for takeoff, landing, and
autopilots.
[0006] GNSS is a navigation system that allows small receivers to
determine their position with respect to the earth using signals
transmitted from satellites. This system typically permits
geo-spatial positioning with world wide coverage. GLS is a system
that combines satellite and ground-based navigation information to
provide aircraft positional information with respect to a
pre-defined approach path during approach, landing and rollout.
[0007] A key issue with GLS is the expected failure modes and
effects of failures in the GLS guidance system. It is anticipated
that the most common failure mode for GLS is the total loss of
signal from satellites for hundreds of seconds. Consequently, a
GNSS/inertial coasting filter has been developed to provide
continuity of service through these outages by continuing to
provide inertially derived position information when the GLS
guidance is unavailable.
[0008] The inertial coasting filter provides aircraft systems with
a reliable backup form of guidance when the GLS guidance signals
are unusable. The guidance system responds as rapidly as possible
to switch to inertial guidance to prevent the inertially based
deviations from becoming corrupted by the errors in the GLS
signals.
[0009] Errors in GLS systems generally occur at a much lower
frequency than errors associated with other guidance sources, such
as instrument landing systems. However, GLS may experience discrete
changes or steps in these steady state errors which occur in
fault-free operation of the system as satellites rise and set.
[0010] The integrity of a GLS guidance system is frequently
specified in terms of an alert limit. An alert limit is a limit on
the maximum allowable GLS error on differentially corrected
deviations transmitted without annunciation of the problem to the
flight crew. However, current standards do not limit the dynamic
behavior of the error when the error is within the alert limits.
This poses a potential problem for the inertial coasting filter
scheme which uses low-frequency information from the nominal GLS
deviations to establish an accurate inertial reference.
[0011] ILS inertial coasting has a coasting duration limitation of
approximately 20 seconds or less which allows adequate time for a
backup station to come on line without requiring the approach to be
aborted. However, given the potential total loss of GLS signal for
hundreds of seconds, to allow an approach to be completed from the
alert height in the presence of such failures, the GLS coasting
duration must be expanded to 60 seconds or more. To accomplish this
safely, more precise inertial coasting is required. Thus, there is
a wider range of detrimental error rates for current GLS than ILS
systems. For example, if the reference error on the GLS deviations
increases at a low enough frequency, the errors will be
incorporated into the blended solution via the inertial bias
estimate, while the conflicting inertial low-frequency information
is rejected. Such a corrupted inertial reference, over time, can
lead to touchdown and rollout off the runway. Such problems lead to
decreased safety for passengers.
[0012] Further complicating the problem is the fact that, due to
the high probability of error steps occurring as a result of normal
satellite configuration changes, such as during satellite rising
and setting, these error events need to be differentiated from
error events that require action to be taken to remove negative
impact on the GLS auto-land performance. Thus, due to the
differences between ILS and GLS, current anomaly detection methods
used in ILS based systems cannot be safely and accurately used for
anomaly detection in current GLS based navigation systems.
[0013] In other words, current GLS landing systems do not employ
coast-skip reset filters or take full advantage of the continuity
provided by coast-skip reset filters. To enable this, it would be
desirable to detect a wide range of ramp errors which are not
detectable by current state of the art anomaly detectors.
SUMMARY
[0014] The different advantageous embodiments provide a computer
implemented method, apparatus, and computer program product for
detecting data anomalies in a landing system. In one embodiment, a
magnitude difference between a blended inertial deviation magnitude
and a raw deviation magnitude is identified to form a magnitude
difference. The magnitude difference is compared to a magnitude
threshold. If the magnitude difference exceeds the magnitude
threshold, an anomaly in the data is detected.
[0015] Another embodiment comprises a computer program product on a
computer usable medium having computer usable program code for
detecting data anomalies in a landing system. The computer program
product includes computer usable program code to identify a
magnitude difference between a blended inertial deviation magnitude
and a raw deviation magnitude in a landing system associated with a
global navigation satellite system to form a magnitude difference;
compare the magnitude difference to a magnitude threshold; and
trigger inertial coasting in response to the magnitude difference
exceeding the magnitude threshold.
[0016] In another embodiment, the apparatus has an anomaly detector
in a landing system associated with a global navigation satellite
system. The anomaly detector includes a coasting filter. The
coasting filter generates a blended inertial deviation magnitude
and a blended inertial deviation rate. A coast skip reset trigger
anomaly detector identifies a magnitude difference between the
blended inertial deviation magnitude and a raw deviation magnitude
to form a magnitude difference and compares the magnitude
difference to a magnitude threshold and detects an anomaly in the
data in response to the magnitude difference exceeding the
magnitude threshold, an anomaly in the data is detected.
[0017] The features, functions, and advantages can be achieved
independently in various embodiments of the present disclosure or
may be combined in yet other embodiments in which further details
can be seen with reference to the following description and
drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
[0018] The novel features believed characteristic of the invention
are set forth in the appended claims. The invention itself,
however, as well as a preferred mode of use, further objectives and
advantages thereof, will best be understood by reference to the
following detailed description of an advantageous embodiment of the
present disclosure when read in conjunction with the accompanying
drawings, wherein:
[0019] FIG. 1 is a diagram illustrating an aircraft manufacturing
and service method in accordance with an advantageous
embodiment;
[0020] FIG. 2 is a diagram of an aircraft in accordance with an
advantageous embodiment;
[0021] FIG. 3 is a diagram of a data processing system in
accordance with an advantageous embodiment;
[0022] FIG. 4 is a diagram illustrating components used for
detecting data anomalies in a landing system in accordance with an
advantageous embodiment;
[0023] FIG. 5 is a diagram illustrating implementation of
coast-skip reset trigger anomaly detection for ramps in accordance
with an advantageous embodiment;
[0024] FIG. 6 is a set of graphs illustrating deviation, rate, and
anomaly flag without rate-detection logic in the presence of a step
in accordance with an advantageous embodiment;
[0025] FIG. 7 is a set of graphs illustrating deviation, rate, and
anomaly flag with rate detection disable logic in the presence of a
step in accordance with an advantageous embodiment;
[0026] FIG. 8 is a diagram illustrating implementation of
coast-skip reset trigger anomaly detection with step detection and
associated rate disable logic to minimize anomalous trips due to
acceptable steps in position resulting from normal satellite
configuration changes in accordance with an advantageous
embodiment;
[0027] FIG. 9 is a diagram illustrating implementation of a slow
ramp detector using inertial velocity input to a coasting filter in
accordance with an advantageous embodiment;
[0028] FIG. 10 is a diagram illustrating implementation of a slow
ramp detector using inertial acceleration input to a coasting
filter in accordance with an advantageous embodiment;
[0029] FIG. 11 is a graph illustrating an error insertion
comparison in accordance with an advantageous embodiment;
[0030] FIG. 12 is a flowchart of a process for detecting data
anomalies in accordance with an advantageous embodiment;
[0031] FIG. 13 is a flowchart of a process for detecting a ramp
error in accordance with an advantageous embodiment;
[0032] FIG. 14 is a flowchart of a process for detecting slow onset
data anomalies using a dynamic threshold in accordance with an
advantageous embodiment; and
[0033] FIG. 15 is a flowchart of a process for utilizing dynamic
thresholds to detect slow onset ramp errors in accordance with an
advantageous embodiment.
DETAILED DESCRIPTION
[0034] Referring more particularly to the drawings, embodiments of
the disclosure may be described in the context of the aircraft
manufacturing and service method as shown in FIG. 1 and the
aircraft as shown in FIG. 2.
[0035] Turning first to FIG. 1, a diagram illustrating an aircraft
manufacturing and service method is depicted in accordance with an
advantageous embodiment. During pre-production, aircraft
manufacturing and service method 100 may include specification and
design 102 of aircraft 200 in FIG. 2 and material procurement 104.
During production, component and subassembly manufacturing 106 and
system integration 108 of aircraft 200 in FIG. 2 takes place.
Thereafter, aircraft 200 in FIG. 2 may go through certification and
delivery 110 in order to be placed in service 112. While in service
by a customer, aircraft 200 in FIG. 2 is scheduled for routine
maintenance and service 114, which may include modification,
reconfiguration, refurbishment, and other maintenance or
service.
[0036] Each of the processes of method 100 may be performed or
carried out by a system integrator, a third party, and/or an
operator as indicated by the "X" in the grid to the right of the
flow diagram of FIG. 1. In these examples, the operator may be a
customer. For the purposes of this description, a system integrator
may include without limitation any number of aircraft manufacturers
and major-system subcontractors; a third party may include without
limitation any number of vendors, subcontractors, and suppliers;
and an operator may be an airline, leasing company, military
entity, service organization, and so on.
[0037] With reference now to FIG. 2, a diagram of an aircraft is
depicted in which an advantageous embodiment may be implemented. In
this example, aircraft 200 has wings 202 and 204 attached to body
206. Aircraft 200 includes wing mounted engine 208, wing mounted
engine 210, and tail 212. Aircraft 200 is produced by aircraft
manufacturing and service method 100.
[0038] Apparatus and methods embodied herein may be employed during
any one or more of the stages of production and service method 100
in FIG. 1. For example, components or subassemblies corresponding
to component and subassembly manufacturing 106 may be fabricated or
manufactured in a manner similar to components or subassemblies
produced while aircraft 200 is in service. Also, one or more
apparatus embodiments, method embodiments, or a combination thereof
may be utilized during production stages for component and
subassembly manufacturing 106 and system integration 108 in FIG. 1,
for example, by substantially expediting assembly of or reducing
the cost of aircraft 200. Similarly, one or more of apparatus
embodiments, method embodiments, or a combination thereof may be
utilized one of during different stages, such as component and
subassembly manufacturing 106, and system integration 108, in
service 112, and/or routine maintenance and service 114 of aircraft
200 in FIG. 2.
[0039] In this illustrative example, aircraft 200 includes a GNSS
landing system (GLS) and an anomaly detector for detecting data
anomalies in the GLS data. The anomaly detector may be implemented
using a computing device for receiving and analyzing signals
received from satellites and/or ground-based stations, such as,
without limitation, the data processing system described in FIG. 3
below.
[0040] Turning now to FIG. 3, a diagram of a data processing system
is depicted in accordance with an advantageous embodiment. In this
illustrative example, data processing system 300 includes
communications fabric 302, which provides communications between
processor unit 304, memory 306, persistent storage 308,
communications unit 310, input/output (I/O) unit 312, and display
314.
[0041] Processor unit 304 serves to execute instructions for
software that may be loaded into memory 306. Processor unit 304 may
be a set of one or more processors or may be a multi-processor
core, depending on the particular implementation. Further,
processor unit 304 may be implemented using one or more
heterogeneous processor systems in which a main processor is
present with secondary processors on a single chip. As another
illustrative example, processor unit 304 may be a symmetric
multiprocessor system containing multiple processors of the same
type.
[0042] Memory 306, in these examples, may be, for example, a random
access memory. Persistent storage 308 may take various forms
depending on the particular implementation. For example, persistent
storage 308 may contain one or more components or devices. For
example, persistent storage 308 may be a hard drive, a flash
memory, a rewritable optical disk, a rewritable magnetic tape, or
some combination of the above. The media used by persistent storage
308 also may be removable. For example, a removable hard drive may
be used for persistent storage 308.
[0043] Communications unit 310, in these examples, provides for
communications with other data processing systems or devices. In
these examples, communications unit 310 is a network interface
card. Communications unit 310 may provide communications through
the use of either or both physical and wireless communications
links.
[0044] Input/output unit 312 allows for input and output of data
with other devices that may be connected to data processing system
300. For example, input/output unit 312 may provide a connection
for user input through a keyboard and mouse. Further, input/output
unit 312 may send output to a printer. Display 314 provides a
mechanism to display information to a user.
[0045] Instructions for the operating system and applications or
programs are located on persistent storage 308. These instructions
may be loaded into memory 306 for execution by processor unit 304.
The processes of the different embodiments may be performed by
processor unit 304 using computer implemented instructions, which
may be located in a memory, such as memory 306. These instructions
are referred to as computer usable program code or computer
readable program code that may be read and executed by a processor
in processor unit 304.
[0046] The computer readable program code may be embodied on
different physical or tangible computer readable media, such as
memory 306 or persistent storage 308.
[0047] Computer usable program code 316 is located in a functional
form on computer readable media 318 and may be loaded onto or
transferred to data processing system 300. Computer usable program
code 316 and computer readable media 318 form computer program
product 320 in these examples. In one example, computer readable
media 318 may be, for example, an optical or magnetic disc that is
inserted or placed into a drive or other device that is part of
persistent storage 308 for transfer onto a storage device, such as
a hard drive that is part of persistent storage 308. Computer
readable media 318 also may take the form of a persistent storage,
such as a hard drive or a flash memory that is connected to data
processing system 300.
[0048] Alternatively, computer usable program code 316 may be
transferred to data processing system 300 from computer readable
media 318 through a communications link to communications unit 310
and/or through a connection to input/output unit 312. The
communications link and/or the connection may be physical or
wireless in the illustrative examples. The computer readable media
also may take the form of non-tangible media, such as
communications links or wireless transmissions containing the
computer readable program code.
[0049] The different components illustrated for data processing
system 300 are not meant to provide architectural limitations to
the manner in which different embodiments may be implemented. The
different illustrative embodiments may be implemented in a data
processing system including components in addition to or in place
of those illustrated for data processing system 300. Other
components shown in FIG. 3 can be varied from the illustrative
examples shown.
[0050] Conventional autopilots with only minor or no reliance on
inertial inputs are sensitive to ILS beam imperfections, such as
beam bends, noise, multi-pathing, and oscillatory behavior of
signals received from satellite and ground stations. Inertial
inputs are provided by an inertial guidance system. An inertial
guidance system is a system that provides the position,
orientation, and velocity of a vehicle by using an inertial
reference frame. The inertial guidance system typically integrates
information gathered from a combination of gyroscopes and
accelerometers to determine the linear and angular accelerations of
the system in the inertial frame.
[0051] Complementary filtering of instrument landing systems (ILS)
position information with inertially-derived acceleration or
velocity results in a very marked attenuation of airplane response
to beam disturbances. The reference complementary filter structure
not only blends inertial information with deviation input, but
additionally enables the deviation input to the complementary
filter to be removed while deviations continue to be provided
and/or derived from the remaining inertial information.
[0052] The filter configuration, with deviation input removed, is
termed inertial coasting. Inertial coasting allows for continuity
in the guidance input to the autopilot during ILS station outages
while the backup station is coming online. Additionally, given the
rapid rate changes associated with unacceptable ILS noise and
interference, the inertially complemented deviations can be
compared with the ILS position information to reject unacceptable
deviation input and trigger inertial coasting to smooth the
autopilot's response to this anomalous guidance. This enhances
passenger comfort through these events and ensures safe guidance
through beam hardovers.
[0053] The use of ILS inertial coasting deviations is limited to
sufficiently span these signal outage and error causing events.
This limit, imposed via annunciation to the flight crew of mode
failure in the effected axis, is in place to eliminate accuracy
concerns related to potential for coasting filter corruption by ILS
beam bends.
[0054] In contrast to ILS, GLS has not been in wide use. GLS can
currently support category I approach and auto-land. A category I
approach and auto-land is a category that permits pilots to land
with a decision height of approximately 200 feet and a forward
visibility of approximately 2400 feet. A simple autopilot is
sufficient for category I. Category II permits pilots to land with
a decision height of approximately 100 feet and a forward
visibility of 1000 feet. At category III, pilots can land with a
decision height as low as 50 feet and a forward visibility of 700
feet. Current GLS systems cannot be used to support category II and
category III operations due to the expected failure modes and
effects of the GLS guidance system.
[0055] The most common failure mode expected for GLS is a total
loss of the signal for hundreds of seconds. A GPS/inertial
"coasting" filter scheme was developed for GLS that is similar to
ILS complementary filter used to provide continuity in guidance
through station outages. The GPS/inertial coasting filter scheme
includes features allowing increased accuracy and, thus, an
extended coasting duration to enable the airplane to continue to
land and rollout after a total loss of GLS guidance below the alert
height.
[0056] The GPS/inertial coasting filter scheme depends on the
premise that the inertial error remains relatively constant through
the duration of the approach. Inertial error, dominated by
accelerometer tilt error conforms to this premise, varying slowly
enough that it can be estimated to a high degree of precision. This
error is estimated in the GLS coasting filters and used to correct
the inertial inputs during coasting operation.
[0057] The different advantageous embodiments recognize that
inertial coasting for GLS requires the GLS ground station to
provide the aircraft systems with information required to determine
with certainty when the GLS guidance signals are unusable. The
guidance system must respond rapidly to switch to inertial guidance
to prevent the inertial signals from becoming corrupted by errors
in the GLS signals. Unfortunately, the ground station cannot
communicate the status of the guidance signals instantaneously, and
therefore, the possibility of corruption still exists. During a
failure condition, it is possible for a differential GPS ground
station to provide corrupted data for up to three (3) seconds
before raising an alarm.
[0058] In other words, the ground station may continue to send a
corrupted signal for up to three seconds before the guidance system
becomes aware of the fact that the signal is corrupted. Thus, the
guidance system will have been using a bad signal for determining
position and location of the aircraft for those three seconds.
[0059] Furthermore, the airplane is allowed to continue to use the
last data provided by the ground station for up to three (3)
seconds after the airborne receiver stops receiving data.
Consequently, there may be a three to six second delay between GPS
signal corruption and detection of the corruption by the airborne
receiver.
[0060] ILS inertial coasting has a coasting duration limitation
varying from approximately 20 seconds down to approximately 4
seconds, depending on altitude. Current navigational systems do not
utilize the coast-skip reset feature. The coast-skip reset feature
of the advantageous embodiments uses the low-frequency information
from the nominal GLS deviations and buffered internal filter states
to free the coasting filter from corruption and establish an
accurate inertial reference.
[0061] In addition, the advantageous embodiments recognize that
nominal low-frequency error characteristics of GLS guidance, in
combination with coast-skip reset features described herein, enable
the accuracy required to remove the coasting duration limitation
imposed on ILS inertial coasting.
[0062] The current GLS coasting filter design is still subject to
misguidance from the corruption of the filter during this three to
six second failure exposure. This potential for corruption of the
filter exists in the currently available ILS/inertial filtering
scheme, however, any negative impact of this corruption is bounded
by the shorter coasting duration required. Coasting from the alert
height through rollout is needed to supplement continuity GLS for
fail-operational auto-land. Therefore, a method to uncorrupt the
filter from an anomalous GLS signal is needed.
[0063] The currently available GPS/inertial skipping filter scheme
allows the coasting filter to recover from up to six (6) seconds of
signal corruption and avoid subsequent misguidance. The
decontamination of the complementary filter is accomplished by
buffering inertial bias error and bias error rate information, held
from sufficiently long ago, to guarantee their integrity has not
been compromised. The corruption-free inertial error and error rate
estimates are then used to provide a correction for the inertial
information so that the sum can be integrated appropriately and a
correct deviation can be produced from the filter even when the GLS
deviation input is removed from the solution.
[0064] The advantageous embodiments recognize that failure-free
errors in GLS occur at a much lower frequency than errors in ILS.
Unique to GLS, however, are the steps in which the errors occur in
fault-free operation, as satellites rise and set. In GPS system
failures, ramps and steps of virtually any magnitude and speed are
feasible but the exposure time should be limited by ground based
augmentation systems (GBAS) time to alert and allowance for missed
messages. Ionosphere gradient anomalies have been found to have
very slowly varying effects but similarly present themselves in the
form of a ramping error.
[0065] The integrity of a GLS guidance system can be specified in
terms of an alert limit which is essentially a limit on the maximum
allowable GLS error on the differentially corrected deviations
transmitted without annunciation. Current standards do not limit
the dynamic behavior of the error when it is within the alert
limits. The advantageous embodiments recognize that this poses a
potential problem for the coast-skip reset, which uses the
low-frequency information from the nominal GLS deviations to
establish an accurate inertial reference. The nominal low-frequency
error characteristics of GLS guidance are what, in combination with
the coast-skip reset, enable the accuracy required to remove the
coasting duration limitation imposed on ILS inertial coasting. If
the reference error on the GLS deviations is increasing at a low
enough frequency, it will be incorporated into the blended solution
via the inertial bias estimate while the conflicting inertial
low-frequency information is rejected. Such a corrupted inertial
reference can, over time, lead to touchdown and rollout off the
runway, which is unacceptable for category II and category III
operations.
[0066] The advantageous embodiments recognize that ramp errors have
the potential to cause noticeable flight path deviations, depending
on the ramp rate of the error applied, even with the ILS anomaly
detection enabled. Even if a coasting filter associated with the
aircraft guidance system is outfitted with coast-skip reset to
provide continuity for signal loss below the alert height and
remove corruption due to ground station detected failures, the
aircraft is still subject to that maximum of six (6) seconds
exposure time to the failure in progress due to the faulty data.
Those six seconds of exposure can result in deviations on the order
of tens to hundreds of feet depending on the nature of the failure.
Deviation of this magnitude from the true glidepath or from runway
centerline while early in the approach is likely not an issue.
However, below the alert height, such a deviation from the desired
path can pose a serious safety concern. In other words, the errors
can cause the aircraft to be hundreds of feet away from the center
of the runway during take-off or landing, which could result in
difficulties during take-off and landing, an inability to take-off
or land, or even collisions with other vehicles or structures.
[0067] The advantageous embodiments recognize that due to the
extension of the inertial coasting duration for GLS approaches,
there is a wider range of detrimental error rates for GLS than ILS.
Further complicating the problem, due to the high probability of
error steps due to normal satellite configuration changes, these
error events must be differentiated from those for which actions
should be taken to remove negative impact on GLS auto-land
performance. Thus, to take full advantage of the continuity
provided by the coast-skip reset filter, it is necessary to detect
a wide range of ramp errors which are not detectable by current
state of the art anomaly detectors.
[0068] Therefore, the different advantageous embodiments provide a
method, apparatus, and computer usable program code for detecting
data anomalies in a landing system utilizing a global navigation
satellite system. Anomalous GLS data may be caused by, but not
limited to, satellite motion, location, satellite configuration
changes, or any fault-free operation as satellites rise and
set.
[0069] In one embodiment, a magnitude difference between a blended
inertial deviation magnitude and a raw deviation magnitude is
identified to form a magnitude difference. The magnitude difference
is compared to a magnitude threshold. If the magnitude difference
exceeds the magnitude threshold, an anomaly in the data is
detected.
[0070] Likewise, a deviation rate is compared to a deviation rate
threshold. If the deviation rate exceeds the deviation rate
threshold and a step detector indicates a ramp has occurred rather
than a step, an anomaly in the data is detected. As used herein,
the deviation magnitude threshold and the deviation rate threshold
are fixed values to which the anomaly detection input is compared
to set a coast status.
[0071] FIG. 4 is a diagram illustrating components used for
detecting data anomalies in a landing system in accordance with an
advantageous embodiment. GLS anomaly detection system 400 is a
system for detecting GLS data anomalies in a GLS system. GLS
anomaly detection system 400 includes an inertial unit 402.
[0072] Inertial unit 402 is an inertial guidance system. Inertial
unit 402 measures linear and angular accelerations applied to the
system in an inertial reference frame to coasting filter 404.
[0073] Coasting filter 404 may also be referred to as an inertially
based deviation backup. Coasting filter 404 may be implemented in
any known or available coasting filter, such as, but not limited
to, an ILS anomaly detector.
[0074] In this example, coasting filter 404 receives inertial data
from inertial unit 402. Coasting filter 404 generates inertially
complemented GLS deviation and deviation rate. The inertially
complemented GLS deviation may also be referred to as a blended
inertial deviation magnitude. The deviation rate may also be
referred to as the blended inertial deviation rate.
[0075] Coasting filter 404 outputs the inertially complemented GLS
deviation and deviation rate to coast-skip reset trigger anomaly
detector 406.
[0076] Coast-skip reset anomaly detector 406 is a new anomaly
detector for detecting data anomalies associated with GLS.
Coast-skip reset anomaly detector 406 compares the inertial
estimates of deviations and rates to GLS deviations and rates to
detect GLS anomalies. In other words, coast-skip reset trigger
anomaly detector 406 compares the inertially complemented GLS
deviation rate output by coasting filter 404 with the raw GLS
deviation. Coast-skip reset trigger anomaly detector 406 also
compares the blended inertial deviation rate with the raw GLS
deviation rate.
[0077] Thus, in this example, currently available coasting filter
404 is supplemented with coast-skip reset trigger anomaly detector
406. Coast-skip reset trigger anomaly detector 406 provides
coasting filter 404 with immunity to step changes in deviation
which are normal for GLS but non-existent for ILS. Coast-skip reset
trigger anomaly detector 406 uses a novel step-detection path which
is discussed in greater detail below. This step-detection path
allows tighter ramp detection thresholds to be set without
contribution to anomalous trip rates.
[0078] Coast-skip reset trigger anomaly detector 406 identifies the
onset of a GLS failure so that inertial coasting can be triggered
before the airplane has responded to the failure in progress. These
failures manifest themselves in the deviations as ramps or steps of
virtually any rate and magnitude. However, steps in the deviations
can also result due to nominal events, such as satellite
configuration changes as satellites rise and set. Since the
probability of these events yielding a position step of less than
or equal to three meters is relatively high (0.9999), it is
desirable that the anomaly detection be able to distinguish between
an anomaly detection trip due to a nominal position step and an
unacceptable step in the position. Coast-skip reset trigger anomaly
detector 406 flags the slowest possible error ramp rate to achieve
the greatest possible coasting performance protection.
[0079] In addition to this basic anomaly detection innovation, an
additional novel anomaly detector, slow ramp detector 410, may
optionally be included to detect ramps that are undetectable by
conventional inertial to deviation input comparison. Slow ramp
detector 410 and coast-skip reset trigger anomaly detector 406
combine to form comprehensive protection against GLS failures which
could cause unacceptable degradation in auto-land performance while
coasting through GLS signal outage. This allows full advantage to
be gained from the continuity supplement in support of category II
and category III certification acceleration and/or operational
approval activities.
[0080] Navigation receiver 408 is a receiver for receiving signals
from satellites and/or ground based navigation stations. The
signals may be lost, distorted, or contain erroneous information
due to a number of causes or problems with the ground station
and/or the satellites.
[0081] However, the consistency with which GPS satellite
aberrations manifest themselves in the GLS deviations makes the
deviations and their associated rates a good choice for monitoring
by coast-skip reset trigger anomaly detector 406 and/or slow ramp
detector 410 to protect coasting touchdown and rollout performance.
This is effective because an error step or ramp in the GLS data
will surface immediately in the raw GLS data but will appear much
more gradually in their inertially complemented counterparts.
Therefore, coast-skip reset trigger anomaly detector compares
inertially complemented GLS deviation and deviation rate output by
coasting filter 404 against their raw GLS counterparts in a given
axis.
[0082] Coast-skip reset trigger anomaly detector 406 measures the
magnitude comparison against a threshold. Coast-skip reset trigger
anomaly detector 406 sets status to coast enable, triggering
inertial coasting, when the threshold is exceeded. Coast-skip reset
trigger anomaly detector 406 measures the rate comparison against a
threshold. Coast-skip reset trigger anomaly detector 406 triggers
inertial coasting based on the comparison when the differences in
deviation magnitude or deviation rate exceed the respective
thresholds. In other words, the rate and magnitude thresholds are
independent and, therefore, the deviation magnitude is compared to
a deviation magnitude threshold while the deviation rate is
compared to an independent deviation rate threshold.
[0083] In other words, the resulting mismatch between the
inertially complemented GLS deviation and deviation rate and the
raw GLS deviation and deviation rate can be measured against a
threshold allowing all potentially hazardous hardovers to be
detected at their onset without relying on ground station
monitoring and/or design of complex monitors against specific
system failure modes.
[0084] Referring now to FIG. 5, a diagram illustrating
implementation of coast-skip reset trigger anomaly detection for
ramps is shown in accordance with an advantageous embodiment.
Coasting filter 500 is a coasting filter, such as coasting filter
404 in FIG. 4. Coasting filter 404 provides blended inertial
deviation to a coast-skip reset trigger anomaly detector for
comparison to the raw GLS deviations 502 for detecting unacceptable
steps in position, such as coast-skip reset trigger anomaly
detector 406 in FIG. 4. An unacceptable step is a step in input
data that should be disregarded, deleted, or otherwise not relied
upon for use in navigational processes due to an anomaly or other
error associated with the input data.
[0085] The coast-skip reset trigger anomaly detector compares
blended inertial deviation to raw GLS deviation difference with a
threshold 504 to identify unacceptable steps in the deviation. If
an unacceptable step is detected, coast-skip reset trigger anomaly
detector sets a status to coast 506 to trigger inertial
coasting.
[0086] Coasting filter 500 also outputs blended inertial deviation
rate to the coast-skip reset trigger anomaly detector for
comparison to the raw GLS deviation rate 508 to detect ramps. The
coast-skip reset trigger anomaly detector compares blended inertial
deviation rate to raw GLS rate difference 508 to deviation rate
threshold 510. If the comparison indicates the threshold is
exceeded and a step detector indicates this is a ramp and not a
step, the coast-skip reset trigger anomaly detector sets a status
to coast 506 to trigger inertial coasting.
[0087] Thus, the coast-skip reset trigger anomaly detector compares
the differenced deviations and rates with a threshold set to
provide protection against the slowest possible GLS error ramp-rate
and against unacceptable steps in GLS position. The two detection
schemes using deviation threshold 504 and deviation rate threshold
510 are designed to work together with the coasting capability of
the complementary coasting filter 500.
[0088] When an error ramp initially onsets, the deviation rate
detection output "b" will trip to trigger the coast-skip reset and
opening the filter to the GLS input. Because the corruption due to
the onset of the anomaly has been removed by the coasting filter
via the coast-skip reset and the GLS deviation is no longer being
input to coasting filter 500, if the ramp error persists, the rate
difference between filtered and unfiltered anomaly detection input
also persists. If the ramp stops and the accumulated error has not
exceeded the deviation magnitude detection threshold, the error is
within an acceptable range and the deviation can be followed again
by the autopilot.
[0089] In this case, the deviation rate detection output "b" will
no longer be true and the deviation detection output "a" will
continue to be false. The coast output clears and allows the GLS
deviation information to again be input to the filters. However, if
the ramp stops but the error exceeds deviation magnitude detection
threshold 504, the output "b" will no longer be true. Because the
error accumulated during the ramp is unacceptable, the deviation
magnitude detection output "a" will sustain the anomaly detection
and the filters continue to coast inertially.
[0090] Thus, the process shown in FIG. 5 allows for successfully
detecting hazardous ramps and steps in GLS data. However,
acceptable size steps may also be rejected in the position solution
due to normal satellite configuration changes. This is primarily
due to the tighter threshold used to catch the wider range of
detrimental error rates for GLS approaches. The negative impact can
be seen by looking at a step in the GLS deviation and comparing the
resulting raw GLS and filtered outputs as they relate to coasting
filter 500. The time history of these outputs and the coast output
anomaly flag from coasting filter 500 are shown in FIG. 6 and FIG.
7 below.
[0091] FIG. 6 is a set of graphs illustrating deviation, rate, and
anomaly flag without step-detection logic in the presence of a step
in accordance with an advantageous embodiment. Graph 602 shows the
raw GLS deviation and inertially blended deviation inputs to the
coast-skip trigger anomaly detector. A three meter step is inserted
at the elapsed time of 100 seconds at 603.
[0092] Graph 604 is a graph of deviation rate input data. Inserting
the three meter step at 603 results in a large jump in the raw GLS
rate input at 605. When differenced with the inertially blended
deviation rate input which initially rejects the step, this
propagates into a large rate difference 508. This large jump occurs
with a step of any magnitude due to the initial rejection of the
step in the filtered rate input due to inertial smoothing. Such a
rate difference will exceed the threshold, setting path "b" in FIG.
5, triggering the filter to coast.
[0093] However, with the filter open to the GLS input, the filtered
and raw GLS deviation-rate begins to track, offset by the magnitude
of the onset step. Because the deviation rates are again tracking,
the path which was holding the anomaly flag is no longer satisfied.
The GLS deviation is re-input to the filter. This is the desired
result because if a step of acceptable magnitude onsets, it can not
be determined if the position was more correct before or after the
step, therefore, the filtered deviation should be allowed to
converge to the new GLS deviation.
[0094] The deficiency of the detector without step detection logic
arises as the filter begins to converge to the new position. For
the filtered deviation to converge to the stepped GLS deviation, a
rate difference is required between the two. Depending on the size
of the step, this has the potential to re-trip the rate detector
and set off a cycle of tripping, un-tripping, and/or attempting to
converge. Graph 606 illustrates an anomaly flag that is repeatedly
tripping and un-tripping from the point where the step is
introduced at the elapsed time of 100 seconds to approximately 105
seconds. As can be seen in graph 606, this repeated cycling of
tripping, un-tripping, and/or attempting to converge can
potentially continue for the duration of the approach, depending on
the size of the step, inertial errors, latency, and other
factors.
[0095] Turning now to FIG. 7, a set of graphs illustrating filtered
and raw GLS deviations, rates, and the coast-skip trigger anomaly
flag with rate detection disable logic in the presence of a step is
shown in accordance with an advantageous embodiment. Again, graph
702 illustrates the inertially blended and raw GLS deviation inputs
to which a step has been introduced at elapsed time 100 seconds.
The step results in a large jump in the GLS rate input in graph
704.
[0096] To eliminate the cycling of the anomaly flag shown in graph
606 in FIG. 6, and ensure that an acceptable sized step is
converged upon by the coasting filter, the characteristic rate-trip
followed by rate-tracking caused by normal and/or acceptable sized
steps in GLS deviation is used to discern a step from a ramp and
appropriately disable the rate detection logic. The step is tracked
to the filter.
[0097] Persistence is required on the tracking of the actual and
filtered rates before the anomaly flag clears to assure the initial
trip rate was truly due to a step in position. This persistence
prevents a step from being declared due to a brief negation of
deviation output caused by momentarily erroneous rate-alignment
between a GLS error ramp and the filtered deviation due to
turbulence or some other outside factor. This additional step
detector and the associated rate-detector disable logic are added
to the original deviation magnitude/rate detector to complete the
coast-skip reset trigger anomaly detector.
[0098] The step detector and associated rate-detector disable logic
are utilized to supplement the coasting filter and coast-skip reset
trigger anomaly detector to eliminate and/or minimize the cycling
of the anomaly flag, as shown in graph 706.
[0099] FIG. 8 is a diagram illustrating implementation of
coast-skip reset trigger anomaly detection with step detection and
associated rate disable logic to minimize anomalous trips due to
acceptable steps in position resulting from normal satellite
configuration changes in accordance with an advantageous
embodiment. Anomaly detector system 800 is a system for detecting
GLS data anomalies using a coast-skip reset trigger anomaly
detector and a trailing edge detector with step detection and
associated rate disable logic to minimize anomalous trips due to
acceptable steps in position resulting from normal satellite
configuration changes.
[0100] Coasting filter 802 is a coasting filter, such as, but not
limited to, coasting filter 404 in FIG. 4. Coasting filter 404 in
FIG. 4 is supplemented by a coast-skip reset trigger anomaly
detector, such as coast skip reset trigger anomaly detector 406 in
FIG. 4.
[0101] Coasting filter 802 outputs blended inertial deviation input
to difference 804 and blended inertial deviation rate input to
difference 814. The coast-skip reset trigger anomaly detector
compares the blended inertial deviation to raw GLS deviation
difference 804 to deviation threshold 806. Deviation threshold
input to comparator 806 is a magnitude threshold.
[0102] If the threshold is exceeded, the coast-skip reset trigger
anomaly detector triggers inertial coasting 810. The coast-skip
reset trigger anomaly detector also compares blended inertial
deviation rate to raw GLS deviation rate difference 814 with
deviation rate threshold 816. If deviation rate threshold 814 is
exceeded and the step detector indicates this is a ramp and not a
step, the coast-skip reset trigger anomaly detector triggers
inertial coasting 810.
[0103] Returning to the example in FIGS. 6 and 7 where a three
meter step is inserted, the effects of the step can be traced as
the effects propagate through anomaly detector system 800. At time
100 seconds when the step is inserted, there is a large difference
observed between blended inertial deviation rate and the raw GLS
rate 814 which exceeds the threshold 816 setting path "b." Because
nothing thus far has reset the input "d", the and-gate of "b" and
"d" is satisfied and coast is set to trigger inertial coasting 810.
At this point, the time histories shown in FIGS. 6 and 7 are
identical.
[0104] However, as the filtered deviation rate once again begins to
track the GLS deviation rate, the added step detection logic for
trailing edge detector 812 becomes active. When blended inertial
deviation rate and the raw GLS rates track persistently, the rate
threshold comparison 816 is no longer exceeded and the rate path of
anomaly detector system 800 "b" is no longer satisfied. Coast is
set to false, re-inserting the GLS deviation into the filter.
[0105] With this additional logic, trailing edge detector 820 of
coast triggers "c" to reset the output "d" to false temporarily
until a time delay has elapsed. During that predetermined period of
time for the time delay, the predicted rate difference sets path
"b" as the filtered deviation converges to the stepped GLS
deviation. Without the step detection logic, this would again cause
the filter to coast. With the step detection logic, however, the
rate-input to the anomaly detection is effectively disabled by the
"d" input to the and-gate. While the rate detection is momentarily
disabled, the performance is still protected by the deviation
magnitude monitor to trigger to coast-skip and activate inertial
coasting via path "a".
[0106] Trailing edge detector 820 enables coasting filter 802 to
have immunity to steps in the position solution due to normal
satellite configuration change, such as, without limitation, steps
large enough to trip the rate detection but small enough that they
do not trip the magnitude detection. Trailing edge detector 820
does not interfere with the rejection of unacceptable sized steps
which trip both the rate detection initially and persistently
exceed deviation threshold 806 at "a". In this example, anomaly
detector system 800 enables both the desired ramp and step error
detection capability and the desired immunity to normal GPS
position steps.
[0107] In this example in FIG. 8, anomaly detector system 800 works
as a step detector and can discern between an anomaly in the data
resolving itself such that no convergence is required, such that
the system remains in coast the whole time, and an acceptable sized
step that anomaly detector system 800 needs to converge to. The
trailing edge of the coast declares a step. It is the trailing edge
of the rate trip coinciding with coast going inactive that permits
the step error to be declared.
[0108] In another embodiment, a similar effect is obtained by
increasing the rate detection thresholds. However, unacceptable
reduction in the ramp detection capability of the detector could
result. This provides protection against GLS aberrations which are
bounded by the time-to-alert.
[0109] A slow ramp detector, such as slow ramp detector 410 in FIG.
4, provides protection against GLS anomalies which are not bounded
by the time-to-alert. These very slowly developing ramp-like
errors, such as, without limitation, those that result from steep
ionospheric gradient, have essentially the same affect on the
touchdown dispersion as a constant guidance error of the size that
the slow ramp error has accumulated at the flare initiation
altitude. Coasting performance, however, is impacted more severely
by these errors. For example, if the reference error is increasing
at a low enough frequency, the error will be incorporated into the
blended solution via the inertial bias estimate, while the
conflicting inertial low-frequency information is rejected. Such a
corrupted inertial reference, over time, can lead to unacceptable
touchdown and rollout off the runway.
[0110] These effects typically cannot be distinguished from the
differences resulting from delays between the inertial data and the
GLS guidance in the presence of turbulence. This motivates the
second innovation in the GLS anomaly detection system, the slow
ramp detector.
[0111] Referring now to FIG. 9, a diagram illustrating
implementation of a slow ramp detector using inertial velocity
input to a coasting filter is depicted in accordance with an
advantageous embodiment. Slow ramp detector system 900 detects ramp
errors which onset too slowly to be detectable by the coast-skip
reset trigger anomaly detector.
[0112] Coasting filter 902 is a coasting filter, such as, without
limitation, coasting filter 404 in FIG. 4. Coasting filter 902
provides inertial velocity error estimates to the slow ramp
detector. The slow ramp detector inputs the inertial error estimate
into a first order lag filter. The error estimate is biased in the
positive and negative directions in the lag filter to create an
upper and lower bound around the error estimate. The upper and
lower bound forms a dynamic threshold that includes both upper
threshold value 910 and lower threshold value 904. In other words,
the upper bound forms upper threshold 910 and the lower bound forms
lower threshold 904. To ensure that the thresholds are initialized
properly, this filtering does not commence until the inertial error
estimates have settled.
[0113] When the inertial error estimate has settled and the
thresholds have been set, the slow ramp detector compares the
inertial error estimate to the upper bound 910 at 912. If the bias
estimate is greater than the upper bound at 912, the slow ramp
detector initiates an annunciation 908 to the flight crew
indicating that extended coasting is unavailable.
[0114] Likewise, the slow ramp detector compares the inertial error
estimate to the lower threshold 904 at 906. If the inertial error
estimate is less than the lower threshold 904, the slow ramp
detector sends annunciation 908 to the flight crew indicating that
extended coasting is unavailable.
[0115] Thus, slow ramp detector system 900 limits the GLS coasting
duration because coasting may be corrupted by the detected ramp. In
this example, the coast-skip reset can not be applied to de-corrupt
the coasting filter because unlike the coast-skip reset detection,
because of the slower rate of the errors detectable by the slow
ramp detector system, there is no guarantee that the ramp has been
detected within the 6 second coast-skip reset buffer window.
[0116] Slow ramp detector system 900 triggers fault response such
that subsequent, persistent loss of GLS deviation information or
trigger of the coast-reset trigger anomaly detector, results in a
mode failure. At slow ramp detection, the loss of extended GLS
coasting capability is annunciated to the flight deck via an
auto-land status annunciation indicating that extended coasting is
unavailable.
[0117] The annunciation indicates the loss of fail-operative
rollout capability. This annunciation allows the pilots to adjust
their minimums and take appropriate action when the decision height
is reached. In one embodiment, the annunciation is a "No Land 3"
annunciation which indicates the loss of fail-operative
capability.
[0118] The exact reason the corrupted coasting filter becomes
detrimental to performance is the principle which allows these
ramps to be detectable by slow ramp detection system 900. For
example, the dominant error on the inertial sources input to
coasting filter 902 is accelerometer tilt error known as the
Schuler error. The behavior of this Schuler error is well known to
oscillate between the minimum and maximum values with an 84 minute
period. Therefore, over the approximately 200 seconds of an
auto-land, there will be little variation in the Schuler error.
[0119] Thus, the inertial error or bias estimate used in coasting
filter 902 is a signal which stays relatively constant throughout
the approach in the presence of fault-free GLS conditions. Coasting
filter 902 also reacts predictably to a slow error in progress by
incorporating the error rate into the bias estimate without
responding to the outside effects of wind conditions, turbulence
and aircraft orientation. This predictable behavior in faulted and
fault-free conditions regardless of outside effects makes is a
convenient choice for monitoring to detect these slow ramp
errors.
[0120] The inertial error can be either positive or negative and
the GLS error can likewise onset in either direction. Constant
inertial bias estimate limits do not provide a stringent enough
bound to enable this magnitude of error to be detectable.
Therefore, the advantageous embodiments recognize that dynamic
inertial error estimate thresholds are needed to create an upper
and lower bound on the bias estimate which can reliably catch error
ramps in either direction, while being able to appropriately drift
with the normal expected drift of inertial error.
[0121] The key to the success of these dynamic inertial error
estimate thresholds is the tuning of the time constant of the lag
filter. It must be fast enough that it allows slow drift of the
limits as the inertial error drifts in its Schuler oscillation
while being slow enough that the bounds do not react to the onset
of a GLS error ramp. The bias estimate reaction to a genuine ramp
error is able to exceed the thresholds, tripping the detector,
before the thresholds have been corrupted by the error ramp.
[0122] Similarly, though the effects of turbulence and wind
conditions are relatively small, the offset applied to the filtered
inertial bias estimate to create upper threshold 910 and lower
threshold 904 on the slow ramp detector is set such that anomalous
trips do not result. This allows slow ramp detector system 900 in
this example, to reliably catch slowly growing GLS errors with
minimal anomalous trips.
[0123] FIG. 10 is a diagram illustrating implementation of a slow
ramp detector using inertial acceleration input to the coasting
filter in accordance with an advantageous embodiment. Slow ramp
detector system 1000 is a coasting filter that comprises a slow
ramp detector.
[0124] Coasting filter 1002 is a coasting filter, such as coasting
filter 404 in FIG. 4. Coasting filter 1002 provides blended
inertial acceleration error estimates. The slow ramp detector
creates an upper and lower bound on the inertial error or bias
estimate which can reliably catch error ramps in either direction
while being able to appropriately drift with the normal expected
drift of inertial error to form dynamic thresholds.
[0125] Slow ramp detector inputs the acceleration bias estimate
into a lag filter and biases the bias estimate in both the positive
and negative directions to create an upper and lower bound around
the bias estimate. The upper bound forms upper dynamic threshold
1010. The lower bound forms lower dynamic threshold 1004.
[0126] The inertial error estimate is compared to upper dynamic
threshold 1010 and lower dynamic threshold 1004. If the subsequent
bias estimate is greater than upper dynamic threshold 1010 at
operation 1012, or less than lower dynamic threshold 1004 at
operation 1006, annunciation 1008 is made indicating that extended
coasting is unavailable. This filtering process, whereby the
inertial error estimate is compared to the dynamic thresholds, does
not commence until the inertial error estimate has settled.
[0127] Referring now to FIG. 11, a graph illustrating an error
insertion comparison is depicted in accordance with an advantageous
embodiment. Graph 1100 shows a sample GLS error insertion
comparison with a negative 0.5 feet per second (ft/s) error ramp
applied at 50 seconds elapsed time in heavy turbulence.
[0128] Line 1102 is the bias estimate with the ramp occurring at
approximately 50 seconds. Lines 1106 and 1108 show the upper and
lower dynamic thresholds associated with the bias estimate with the
ramp that is illustrated in line 1102. For comparison, line 1104
shows a nominal bias estimate with no anomalies. Lines 1110 and
1112 show the nominal bias estimate's associated dynamic thresholds
with no GLS errors inserted.
[0129] At approximately 35 seconds of elapsed time, the criteria
have been met in the approach to enable this detector. Then, at
approximately 50 seconds of elapsed time, the GLS error ramp is
inserted. Regardless of GLS latency and turbulence applied, the
insertion of the 0.5 ft/s GLS error ramp is attributed to the
inertial error and drives the bias estimate that is being monitored
to violate the dynamic thresholds.
[0130] The drift of the dynamic thresholds in lines 1106 and 1108
is slow enough that they can accurately detect such a ramp error
even though they begin to drift in the ramp's direction at its
onset.
[0131] FIG. 12 is a flowchart of a process for detecting data
anomalies in accordance with an advantageous embodiment. The
process in FIG. 12 may be implemented by an anomaly detector using
a magnitude threshold, such as coast skip reset trigger anomaly
detector, such as coast skip reset trigger anomaly detector 406 in
FIG. 4.
[0132] The process begins by receiving a blended inertial deviation
magnitude from a coasting filter, such as coasting filter 404 in
FIG. 4 (operation 1202). The process compares the blended inertial
deviation magnitude against the raw GLS magnitude to identify a
magnitude difference (operation 1204).
[0133] The process makes a determination as to whether the
magnitude difference exceeds a magnitude threshold (operation
1206). If the magnitude difference does not exceed the threshold,
the process terminates thereafter. Returning to operation 1206, if
the magnitude difference does exceed the threshold, the process
triggers inertial coasting (operation 1208) with the process
terminating thereafter.
[0134] In one embodiment, triggering the inertial coasting also
triggers de-corruption of the coasting filter via the coast-skip
reset. The coast skip reset de-corrupts the coasting filter by
removing the erroneous blended inertial deviation data from the
filter.
[0135] Referring now to FIG. 13, a flowchart of a process for
detecting a ramp error is shown in accordance with an advantageous
embodiment. The process in FIG. 13 may be implemented by an anomaly
detector using a deviation rate threshold, such as coast skip reset
trigger anomaly detector, such as coast skip reset trigger anomaly
detector 406 in FIG. 4.
[0136] The process begins by receiving a blended inertial deviation
rate from a coasting filter (operation 1302). The process measures
a difference between the blended inertial deviation rate and the
raw GLS rate (operation 1304). The process then makes a
determination as to whether the rate difference exceeds a rate
threshold (operation 1306). If the difference does not exceed the
rate threshold, the process terminates thereafter.
[0137] Returning to operation 1306, if the difference does exceed
the rate threshold, the process determines whether the step
indicator indicates this is a ramp rather than a step (operation
1308). If this is a step, the process terminates thereafter.
[0138] Returning to operation 1308, if this is a ramp, the process
triggers inertial coasting (operation 1310) with the process
terminating thereafter. In other words, if the difference exceeds
the threshold and this is a ramp and not a step, inertial coasting
is triggered and de-corruption of the coasting filter is
initiated.
[0139] FIG. 14 is a flowchart of a process for detecting slow onset
data anomalies using a dynamic threshold in accordance with an
advantageous embodiment. The process in FIG. 14 may be implemented
by an anomaly detector using a slow ramp detector, such slow ramp
detector 408 in FIG. 4.
[0140] The process begins by making a determination as to whether
the inertial error estimate is settled (operation 1402). If the
error estimate is not settled, the process returns to operation
1402 until the error estimate is settled. The process then inputs
the inertial error estimate into a first order lag filter
(operation 1404). The process biases the error estimate in the
positive and negative direction to establish an upper and lower
bound around the inertial error estimate to form a dynamic upper
threshold and a dynamic lower threshold (operation 1406).
[0141] The process compares the un-lagged inertial error estimate
to the dynamic thresholds (operation 1408). The process makes a
determination as to whether a ramp error is detected (operation
1410). A ramp error is detected if the un-lagged error estimate is
greater than the dynamic upper threshold or less than the dynamic
lower threshold. If a ramp error is not detected, the process
terminates thereafter.
[0142] Returning to operation 1410, if a ramp error is detected,
the process sends an annunciation indicating the loss of extended
GLS coasting capability (operation 1412).
[0143] Turning now to FIG. 15, a flowchart of a process for
utilizing dynamic thresholds to detect slow onset ramp errors is
shown in accordance with an advantageous embodiment. The process in
FIG. 15 is a more detailed description of operation 1410 in FIG.
14. The process may be implemented by a slow ramp detector, such as
slow ramp detector 410 in FIG. 4.
[0144] The process begins by making a determination as to whether
the un-lagged inertial error estimate, otherwise referred to as the
bias estimate, exceeds an upper dynamic threshold (operation 1502).
If the bias estimate does not exceed the upper threshold, the
process makes a determination as to whether the bias estimate is
less than a lower dynamic threshold (operation 1504). If the bias
estimate is not lower than the lower threshold, the process
terminates thereafter.
[0145] If the bias estimate is higher than the upper threshold at
operation 1502 and/or if the bias estimate is lower than the lower
dynamic threshold at operation 1504, the process detects a ramp
error (operation 1506) with the process terminating thereafter.
[0146] The different advantageous embodiments provide a computer
implemented method, apparatus, and computer usable program code
detecting data anomalies in a landing system. In one embodiment, a
magnitude difference between a blended inertial deviation magnitude
and a raw deviation magnitude is identified to form a magnitude
difference. The magnitude difference is compared to a magnitude
threshold. If the magnitude difference exceeds the magnitude
threshold, an anomaly in the data is detected.
[0147] The process is related to GNSS landing systems (GLS) and
more particularly, to detecting anomalous GLS data by comparing raw
GLS data magnitudes/rates with inertially blended data and applying
a threshold to the differences. Delay circuits are also utilized to
implement a coast-reset-trigger anomaly detector, as well as a slow
ramp detector to detect unexpected deviations in GLS data.
[0148] The process provides protection against GLS failures which
could cause unacceptable degradation of data while coasting through
GLS signal outage by ignoring step changes in deviation due
fault-free operations of satellites as they rise and set, as well
as other normal satellite configuration changes. This process
allows tighter ramp thresholds to be generated and utilized for
anomaly detection. The process detects slow ramps that are
undetectable by conventional comparisons of inertial data to
deviation input.
[0149] Thus, the process provides a coast-skip-reset anomaly
detection method for triggering coast mode while ignoring normal
steps in satellite position error. The process also provides a slow
ramp mode for detecting slow degradation of satellite data. This
provides additional capability for earlier detection of anomalous
behavior in GPS signals. When the loss of extended GLS coasting
capability is detected, an annunciation is made to the flight deck
indicating a loss of fail-operative rollout capability. This is an
improvement over existing landing systems and solves a problem with
current GLS systems. This system offers customers additional
confidence in the navigational accuracy. This system increases
safety by ensuring the integrity of the coasting filter deviations
as a backup guidance source. In other words, the system increases
confidence in the navigational accuracy of the system and ensures
accuracy of the backup guidance system.
[0150] The advantageous embodiments provide a safe category I,
category II, and even category III GLS systems. The process may be
utilized to detect anomalous GLS data in GLS systems for navigation
during take-off, autopilot, navigation, and/or landing associated
with any type of aircraft or navigable device, including, but not
limited to, airplanes, unmanned aerial vehicles (UAV), air or space
vehicles, rockets, driver-less vehicles, such as driver-less cars,
and/or any other type of manned or unmanned aircraft.
[0151] The flowcharts and block diagrams in the different depicted
embodiments illustrate the architecture, functionality and
operation of some possible implementations of apparatus, methods
and computer program products. In this regard, each block in the
flowchart or block diagrams may represent a module, segment, or
portion of computer usable or readable program code, which
comprises one or more executable instructions for implementing the
specified function or functions. In some alternative
implementations, the function or functions noted in the block may
occur out of the order noted in the figures. For example, in some
cases, two blocks shown in succession may be executed substantially
concurrently, or the blocks may sometimes be executed in the
reverse order, depending upon the functionality involved.
[0152] The description of the different advantageous embodiments
has been presented for purposes of illustration and description,
and is not intended to be exhaustive or limited to the invention in
the form disclosed. Many modifications and variations will be
apparent to those of ordinary skill in the art. Further, different
advantageous embodiments may provide different advantages as
compared to other advantageous embodiments. The embodiment or
embodiments selected are chosen and described in order to best
explain the principles of the invention, the practical application,
and to enable others of ordinary skill in the art to understand the
invention for various embodiments with various modifications as are
suited to the particular use contemplated.
* * * * *